• Welcome to Autism Forums, a friendly forum to discuss Aspergers Syndrome, Autism, High Functioning Autism and related conditions.

    Your voice is missing! You will need to register to get access to the following site features:
    • Reply to discussions and create your own threads.
    • Our modern chat room. No add-ons or extensions required, just login and start chatting!
    • Private Member only forums for more serious discussions that you may wish to not have guests or search engines access to.
    • Your very own blog. Write about anything you like on your own individual blog.

    We hope to see you as a part of our community soon! Please also check us out @ https://www.twitter.com/aspiescentral

How to tell if the audio has been altered in a video?

Greatshield17

Claritas Prayer Group#9435
Hello, I have really quick question here, how can tell if audio, specifically a person's voice, has been altered in a video? I can't show the video with you all here, but does anyone know how to test or figure out if a voice has been altered?
 
Doesn't look like a process for amateurs. I can read waveforms to a limited extent, but have never considered what to look for between an authentic sound source and one suspected of tampering. Unless of course one source is digital and the other is analogue. I suppose it would help to be able to compare one waveform thought to be genuine compared to one thought to be a fake. (I do occasional sound editing of music with Audacity.)

 
Last edited:
In what way do you think the voice has been altered? There are some ways you might be able to find inconsistencies. If you are looking for chopped up audio you can usually tell this by running the audio through a sonograph but unless you know what you are looking for the process can still be a bit mysterious.

I have some experience with sound engineering. You can't always be sure if audio has been modified. You can maybe see little "tells" that would perhaps indicate editing.

If you want to know if a voice has been altered in some way though, this would require lots of reference material that you can be sure has not been altered in any way.

For example, if you think the pitch of a voice has been altered, this will modify the formants. When a human speaks, the mouth acts as a filter. If you look at this in a spectrum analyser you can see the same shape moving up and down the range of harmonics, but if it has been pitch shifted then this shape will spread out or stretch the higher it's pitched. It's harmonic composition will be in a different ratio. Some people call this the "Pinkey and Perky" effect but I've also heard it called "chipmunking".

I guess I would say, it's not an exact science and prone to error, particularly if you don't have a frame of reference for the pure unmodified voice.
 
You also have to consider the medium in which is playing/broadcasting the audio. Considerations like pitch, wow and flutter can skew the way something- or someone should sound, versus how it is being played.

Reminds me of people with non direct-drive turntables. Where the big and thick rubber band that drives the turntable begins to wear out, and while you may or may not initially hear it, you can see it from how the strobe light no longer functions to indicate optimal pitch, wow and flutter. And eventually human voices start to sound awful.

And some online services deliberately increase speed to conserve bandwidth...like YouTube. Using an algorithm they call "time stretching", it's not supposed to alter the pitch. But then who knows about the audio integrity of the original recording itself? Which might simply reflect an unintentionally flawed recording.

In essence, there are many variables in the equation you are pondering.
 
Last edited:
In what way do you think the voice has been altered? There are some ways you might be able to find inconsistencies. If you are looking for chopped up audio you can usually tell this by running the audio through a sonograph but unless you know what you are looking for the process can still be a bit mysterious.

I have some experience with sound engineering. You can't always be sure if audio has been modified. You can maybe see little "tells" that would perhaps indicate editing.

If you want to know if a voice has been altered in some way though, this would require lots of reference material that you can be sure has not been altered in any way.

For example, if you think the pitch of a voice has been altered, this will modify the formants. When a human speaks, the mouth acts as a filter. If you look at this in a spectrum analyser you can see the same shape moving up and down the range of harmonics, but if it has been pitch shifted then this shape will spread out or stretch the higher it's pitched. It's harmonic composition will be in a different ratio. Some people call this the "Pinkey and Perky" effect but I've also heard it called "chipmunking".

I guess I would say, it's not an exact science and prone to error, particularly if you don't have a frame of reference for the pure unmodified voice.

Is it speech or singing you are looking at?
It's speech and I suspect the pitch and tone have been altered; I'll look into what you have suspect, thanks!
 
Experience with autotune, melodyne, formant filters and common DSP / pitch-shifting algorithms can go a long way, as well as reverse-engineering to see if the output is more normal and natural-sounding. I can help you, if need be.
 

New Threads

Top Bottom